Media Release: National poll shows ABS must change flawed religion question next Census

A pair of national Essential Research polls carried out during July again validate ongoing claims that the Census question on religious affiliation overstates the degree of religiosity in Australia, leading to undue funding and privilege for religious interests at the expense of public institutions.

A big thank-you to all those supporters who contributed to our poll fundraising campaign, which helped finance these surveys!

“How much evidence do we need? The results of our national Essential Research polls, released this week, validate recent academic research that Census Question 23 misleads the public,” says National Secular Lobby (NSL) Ambassador, Associate Professor Paul Willis.

“The Census question on religion collects data which is distorted and unfit for being used as the basis for policy and funding decisions,” he says.

“The big question is why the Australian Bureau of Statistics (ABS) is unwilling to take action to fix it.”

Evidence indicates that religiosity in Australia is greatly overstated, meaning additional taxpayer funds worth billions are granted to private religious businesses in education, health, and aged care.

The initial problem is one of psychology. A ‘leading’ question will influence how people interpret and answer it. The ABS’s approach of asking for a person’s religion without first asking people whether they have one will suggest to some that they are expected to respond with a religion, even if it’s one they were only exposed to in childhood and with which they do not currently identify.

Two national Essential Research polls run by the NSL during July provide further evidence of this problem.

The first survey asked the same question that the ABS poses in the Census: “What is your religion?” This attempted to model the change in the non-religious cohort over the five years since the previous census.

This question resulted in a 'No religion' figure of 41 percent and a religious figure of 56 percent.

The second asked a two-part, non-leading question instead: firstly “Do you currently have a religion?” and then going on to gather a specific denomination.

The non-leading alternative question resulted in a ‘No religion’ figure of 52 percent and a religious figure of 44 percent.

In the 2016 Census, the ABS reported a ‘No religion’ figure of just 30 percent and a religious figure of 60 percent.

Full statistics and data on the surveys are available here.

Even allowing for an expected increase in the ‘No Religion’ demographic over the last five years (modelled with the first poll), removing the leading nature of the question resulted in a highly significant difference.

“From these statistics, it’s apparent that the census question leads people towards supplying particular responses by implying that everyone has a religion, even if it’s only a loose cultural or historical affiliation,” says Associate Professor Willis.

“The inflated religiosity figures allow the federal government to justify granting billions of dollars in funding to private religious institutions at the expense of cash-strapped public schools and hospitals.”

“Public education is grossly under-funded and under-resourced by governments. Like healthcare, aged care and welfare, education is being outsourced by governments to private religious businesses, who are free to funnel tax-free profits to their religious masters.”

“It begs the question whether these inflated census figures underpin an ongoing conservative agenda to fund more religious schools, a claim made against PM Malcolm Turnbull by the Australian Education Union and evidenced in an analysis of education funding under Scott Morrison,” Associate Professor Willis says.

“The ABS states that the current data collection practice “aligns with delivery and planning needs for religious organisations” – so the question has been designed specifically with religious service providers in mind.”

Aside from the leading nature of the question, the NSL points out several other problems:

  • The Census and Statistics Regulation 2016, section 9, states that the ABS must collect a “religion or religious denomination,” not a “religious affiliation.”
  • The ABS has conceded to the NSL that a lapsed family religion or even a response based on “the core values that a person aligns themselves to” may be included as a religion, further reducing the accuracy and value of the data.
  • Listing “Atheism” as an example of religions that may be specified using an “Other” response contravenes the High Court definition of a religion (item 3.3), framed in 1983.

“With a $670 million price tag, you would hope the data collected by the census would be meaningful and accurate, but for at least one question, it won’t be,” says Associate Professor Willis.

“All politicians and media would do well to read the new academic research paper, Religiosity in Australia, to better understand the full extent of these problems.”

“The report explains how religious influence has come to dominate the social and political spheres -- not from any popular base but from a narrow and highly evangelical band of religious conservatives.”

“These two national Essential Research polls from the NSL verify the exhaustive academic research referred to above – all of which illustrates the need to correct this ongoing census bias,” Associate Professor Willis said.


The NSL is one of a number of Australian pro-secular groups currently involved in the Census21 campaign, aiming to improve the accuracy of religion statistics in the Census.